Thanks Björn for this very clear and instructive article.

I agree with Dwight, the ommission of the 1:1 format is a mistake.

björn, thanks to your illustration i finally understand how the aspect ratio change works.

Wonder if Panasonic didn't just shrink the image circle to make the concept work on the sensor - that may explain how they squeezed an f2 lens into where they had an f2.8 before!

1:1 is included in firmware 2.0

http://www.panasonic.co.uk/html/en_GB/News/Latest...

Note: For the 4:3 aspect ratio (at the reported 2.0 micron/pixel pitch) - as a result of the active area being smaller than the full (potentially active) sensor size, the specified "1/1.63 inch" diagonal is "shaved" down to "1/1.76 inch diagonal" size. The corresponding (diagonal length) is 9.12 mm ( 7.300 mm wide, 5.475 mm high ).

This is (essentially) the same (4:3 aspect ratio) area of the image sensor in the FZ30 and FZ50 (specified as 1/1.80 inch).

(Slightly off topic), but relevant to a more comprehensive comparison between theLX3 and the FZ30/50 (which all have similar pixel densities on essentially the same sized active sensor area):

The shorter focal length of the LX3 (5.1 mm, as opposed to the 7.37 mm of the FZs) yields double the Depth of Field that the FZs do (at a subject distance under about 2 Feet). Similarly, (at F=2.8) the Depth of Field equals the Subject Distance for an LX3 at about half the distance as that of the FZs, and reaches "Infinity" at a lower Subject Distance than that of the FZs.

Too bad the ".RW2" (12 MB) files appear to apply "Pre-RAW" NR" throwing out low-level contrast information, yielding "RAW" data (that appears by observation to be) inferior to the ".RAW" files made by the FZs (20 MB in file size), where it comes to the details. While the FZs have a larger, higher quality Leica lens system, a price appears to clearly be paid by this file-size compression occurring in the generation of the ".RW2" files by the LX3 (and other Lumix cameras generating ".RW2" files).

Corrections to my previous post - I spoke too soon ...

One (small) detail is that Panasonic states that the pixels on the image sensor in the DMC-LX3 have 2.025 x 10^(-6) Meter size (as opposed to the 2.0 x 10^(-6) meter size reported on the diagram at:
http://www.pbase.com/viztyger/test
A slight (but not large) difference.

Upon further reflection, I find it hard to determine the *exact* active sensor size of the 4:3 (or the two other) aspect ratios. Nevertheless, it appears to be true that it is a value not far off that of the DMC-FZ30/50 image sensor size (1/1.8 Inches).

The spacing between the pixels themselves is unclear and unknown to me. The 1/1.8 Inch image sensor on the FZ30/50 has an active sensor (height) of 5.319 mm (as stated by seemingly official specifications published). 2736 pixels multiplied by 2.025 x 10^(-6) Meters would equal 5.5404 mm.

However, a 1/1.63 Inches diagonal dimension equals 15.5828 mm (whereas the diagonal dimension corresponding to 5.5404 mm height with a 4:3 aspect ratio should equal 9.234 mm ... ). I am quite honestly confused by this seeming incongruity (between 15.5828 mm and 9.234 mm). I reason that the individual pixels must have some inter-pixel spacing that is larger than the individual pixel size (thus accounting for the seeming incongruity). Please post here if you know more about the specifics as to the (actual) dimensions of the (actual) active height and width of the 4:3 aspect ratio area of this DMC-LX3 image sensor.

Bigger always being better in the minds of marketeers and the consuming audience, it appears that these diagonal dimensions (i.e., 1/1.63", 1/1.8", etc.) represent the maximum physical dimension that can be boasted - that of the outer reaches of the entire physical package.

It appears that the active diagonal dimension (for the utilized 4:3 aspect ratio portion of the 1/1.63 inch sensor) is, indeed 9.234 mm ( height=5.5404 mm, width=7.38720 mm ). The ratio between the diagonal of the active region, and the outer edges of the entire physical assembly, equals about 60%. That's about 20% unused between the active portion and the physical edge of the ehtire assembly. If one peers at the photograph of the entire item itself at:
http://www.panasonic.net/avc/lumix/popup/lx3_inte...
it appears that such is (roughly) the case.

The pixel pitch is assumed to be 2.025 x 10^(-6) Meter in deriving the dimensions listed above. The business about the pixel pitch being 2.025 microns (as opposed to the 2.0 micron figure relied upon in the graphic above in this original post
derives from the "CCD Performance is higher, even though it's about the same size, and has the same number of effective pixels" section at:
http://www.panasonic.net/avc/lumix/popup/lx3_inte...

Note: The claim in the text of the above cited section - that a mere 1.25% increase in pixel pitch *magically* resulted in an astounding 20% in efficiency - reeks of pure marketing department "snake-oil". Feel-good fantasy for the sales floor to have a wet-dream over. It is no less than (about) 8 times more than what a physicist would/could accurately report. If it is the case that a 20% efficiency increase resulted from the implementation of this (new) sensor, it may well likely actually derive from other parameters - such as (perhaps) an increased signal (or just an increased signal/noise ratio, allowing higher levels of post-pixel "gain") resulting from the improvements gushed-over in the "Repositioning circuits with the priority on image quality" section. Or, it may derive (in part) from other changes in the construction/materials of the photo-sensors themselves.

Whatever it is, they sure aren't going to tell you ...

Remember, advertising is all about engendering the fuzzy emotional ooze and physical spacey tinkles of a bordello for the consumer market. Even if the mere 1.25% difference in pixel-pitch represents (as one might suspect) nothing "special" (or "miraculous"), the (rhetorical) "marketing department" will weave a tall tale to capture your imagination (and disposable income).

Remember, the LX3 (in the hands of DxO Labs testing) seems to sport (about) 35% lower (actual) ISO than the camera (and Panasonic) would have you believe:
http://www.dxomark.com/index.php/eng/Image-Qualit...
Note: click on the "ISO Sensitivity" tab to bring up the graphed results (pop-up data will display the exact measurement values upon cursor "mouse-over").

Few things in the "consumer state" are what they appear to be ... :)

Nevertheless, they got my money (too), and (despite the proprietary, non-straightforward hocus-focus surrounding the marketing world), I'm liking what I am holding in my paws. The LX3 is (nearly) what a compact 9-ounce version of FZ50 *could* have been (were it not for it's wretched Venus III JPGs). Gamma and sharpening (even with the Sharpness at minimum of -2) are a bit much in the JGPs - the (allegedly 12-bit depth) 12 MB ".RW2" "baby" RAWs are pretty good, too.

They have (probably) squeezed the most out of the Venus IV Engine with the LX3 ... ?

The Venus V (better known as Venus HD) Engine (on the pixel-overstuffed, smaller-sensor FX-48, anyway) shows no evidence so far (perhaps thanks to the marketing-department-driven design factors) of doing better, itself. Venus V seems more about HD video (dual processors, and all) than about moving towards the quality that the LX3 thankfully makes more possible (due to it's adjust-ability as well as it's glass, sensor, hardware, and firmware, etc.).

Regarding the claims made at the Panasonic web-page:
http://www.panasonic.net/avc/lumix/popup/lx3_inte...

it would seem that the around 34.26% increase in light gathering alone when comparing the (4:3 aspect ratio) use of the LX3 image sensor (as opposed to the LX2 image senso)r buys Panasonic 2.5587 dB of the 3.0836 dB (42.62%) improvement that they claim overall between the two individual image sensors.

Yet, they only attribute 1.5 dB (18.85%) of the claimed 3.0836 dB (42.62%) total improvement to increased light-gathering abilities, whereas it (actually) would seem (instead) to account for 2.56 dB (34.28%) of that 3.0836 dB (42.62%) total claimed improvement.

Interestingly, that leaves a mere 0.5249 dB (6.23%) of "miracles" to emerge as a result of "pixel-level-magic" ...

The 1.25% increase in pixel-pitch accounts for 0.2158 dB (2.52%) of the above - leaving only a remainder of only 0.3091 db (3.63%) that would seem to be able to be genuinely attributed to "pixel-structure-magic" as a result of their implicit "groovy-ness" ...

This, coupled with reported 34% lower (actual, tested) ISO than the LX3 reports when used (see "IS0 Sensitivity" chart at:
http://www.dxomark.com/index.php/eng/Image-Qualit... )
constitutes a fair amount of hokey claims. Shame on you, Panasonic Marketing Department ...

The BS appears to run pretty thick and deep from Panny, it seems. But sizzle sells much faster than steak, and few see through the veils. However, if my company lost $4 Billion in 2008, I might be compelled to push the envelope of credibility, too.

It's a good thing that (despite the rhetorical convolutions in marketing) the LX3 is a *decent* camera. Wouldn't it be nice if that (and not pretty fibs and sweet nothings) was what mattered in the marketplace? Don't hold your breath ... ;-)